Observational Designs Flashcards
What is exploratory research?
Serves as a starting point to generate questions.
Useful for insight that leads to scientific inquiry.
Does NOT generate conclusive or generalizable evidence.
Can understand prevalence of problem.
Can understand correlation.
What is conclusive research?
Provides evidence that can be used to help reach conclusions.
Often association with evaluating interventions.
What goals can conclusive research have?
Document scope.
Test hypothesis.
Identify sequelae.
Evaluate interventions.
What is a research design?
Strategy of study to answer a question.
What is prevalance?
Refers to total number of people who currently have condition.
What is incidence?
Refers to the number of people newly diagnosed with the condition.
What is descriptive research?
Answers the questions of who, what, where, when, and how?
What is causal research?
Provides information determining CAUSE of the health issue and evaluates effectiveness of interventions.
What is qualitative research?
Involves collecting data that is not numerical.
Observational.
Aligns with exploratory.
What is a theme?
Abstract construct that investigators identify before, during, or after data collection.
What is quantitative research?
Involves collecting information based on numbers.
PPT responses transformed into numerical values and statistical procedures are employed for data analysis.
What are some examples of quantitative methods?
Cross-sectional, successive independent samples, longitudinal, case-control, case-crossover.
What is an experimental study?
Involves manipulation and randomization.
Tries to control extraneous factors.
What is an observational study?
Involves no interfering or manipulation. Measured items but not assigned.
How is causality determined from a study?
Study design, not statistical methodology, determines how the data can be used.
Ability to imply causality comes from study design.
Statistical methodology can help to control/help improve validity of conclusion and recognize association between X and Y is not just because of a certain thing.
Describe a cross-sectional study.
Occur at fixed time point and data collected only once per individual.
Exposure and outcome measured to calculate prevalence, typically.
Prevalence is sample with characteristic in population/total number of people in sample.
What can a cross-sectional study determine?
5 things:
Prevalence. Population characteristics. Number with and without disease. Difference among people. Relationships among variables for a given population.
What are strengths of cross-sectional study?
5 strengths:
Measure prevalence.
Proportion of persons in a population with a particular disease / total population at risk.
Quick and inexpensive.
Can study multiple outcomes.
High PPT rate and no attrition.
More ethically valid.
What are weaknesses of cross-sectional study design?
7 weaknesses:
Lack of causality.
Difficult to use for rare outcome.
Incidence cannot be determined.
Unable to determine temporal order or directionality.
Does not control for cohort effects.
Third variable problem.
Selection bias.
What is a cohort effect?
Group shares characteristics within defined time period and can affect effects.
What is third variable problem?
Some other variable that is not taken into account that affects independent variable and dependent variable.
Makes it look like they are correlated.
What is selection bias?
Distortion of measure of association due to sample selection that does not reflect the target population.
What is a successive independent samples design?
Simple variant of cross-sectional design.
Difference is that it uses a series of cross-sectional studies to detect changes or TRENDS in a population over time.
Same population must be sampled each wave.
What is weakness of successive independent samples?
3 weaknesses:
BH, perception, attitude trends hard to interpret.
Cause of change not identified.
Not a within-person change.
What is a case-control study?
PPTs are chosen on the basis of if they have or do not have the disease.
The exposure histories of cases and controls are compared with respect to the disease outcome retrospectively.
Helpful to determine relative importance of predictor variable in relation to presence of absence of disease.
What are strengths of case-control study?
3 strengths:
Helpful when looking at rare conditions.
Good with limited cost.
Allows demonstration of RL with limited data.
What are weaknesses of case-control study?
5 weaknesses:
Focus on only one outcome.
Prone to sampling bias.
Observational bias on part of assessor (if they know someone if case versus control that could bias their recording of the data).
Recall bias on part of PPT.
Ascertainment bias on both part of researcher and PPT.
What is a cohort study?
Occurs over time rather than at fixed time points. Data collected on PPTs at multiple points in time.
What are strengths of cohort study?
6 strengths:
Can determine changes in health-related outcomes between persons and within persons.
Can look at relative risk ratios.
Establishes temporal order of risk and outcome.
Predictive ability of risk and protective factors.
Capture incidence.
Captures multiple outcomes.
What are weaknesses of cohort study?
4 weaknesses:
Not suitable when outcome is rare.
Time consuming and expensive.
Large sample needed.
Attrition is high.
What is a cohort sequential study?
Combination of cohort and cross-sectional study. Cohort followed to track risk factors and health outcomes over time. Additionally, cohorts added at a later time.
What are strengths of cohort-sequential study?
3:
Can compare across cohorts.
Changes within PPTs.
Changes between PPts.
What are weaknesses of cohort-sequential study?
3 weaknesses:
Expensive.
Attrition.
Poor casuational analysis.
What is a case crossover study?
Cases serve as their own controls (no controls are recruited).
What are strengths of case crossover?
6 strengths:
For exposures that are difficult to measure or with rapid onset.
Situations where risk factors are hard to match.
Eliminates between-person confounding
Multiple control windows.
Fewer subjects.
See within person changes and see if associated with certain health outcomes.
What are weaknesses of case crossover study?
2 weaknesses:
Recall bias on part of PPT.
Requires careful selection for length and timings of windows (case exposure window and control window timings).
What are the two types of case crossover designs?
Pair-matched interval approach:
Risk exposure period (case exposure window) is matched to an earlier time period called the control window. Compare case exposure window to control window. Case exposure is where something has changed.
Usual frequency approach:
Exposure during case exposure window is compared with the predicted exposure based on typical exposure pattern preceding the outcome.
What are some biases in cross-sectional studies?
4:
Non-response bias
Ascertainment bias
Recall bias
Sampling bias
What are some biases in case-control studies?
5:
Selection bias Recall bias Confounding Ascertainment bias Observer bias
What are some biases in cohort studies?
6:
Non-response bias Healthy entrant bias Attrition Information bias Response bias Allocation bias
What are some biases in case-crossover study?
2:
Selection bias
Confounding
What are some biases in cohort-sequential studies?
Recall bias
What is allocation bias?
Occurs when there is a systematic difference in how PPTs are assigned groups in a trial.
What is ascertainment bias?
Ascertainment bias generally refers to situations in which the way data is collected is more likely to include some members of a population than others. Systematically exclude some ppts from final results.
What is attrition bias?
Refers to unequal loss of PPTs from different groups in a trial.
What is response bias?
Response bias happens once people are in the study (so response bias is not selection bias). Response bias is when participants’ responses to a study assessment are systematically different from their actual experiences (e.g., people often say they drink less alcohol than they actually do).
What is non-response bias?
Non-response bias can result in selection bias. Non-response bias occurs when people who agree to participate in a study are systematically different from those who do. For example, if you invite people to take a survey and those who accept vs. decline your invitation are systematically different from one another (e.g., if women are more likely than men to accept the survey invitation), your sample is biased
What is recall bias?
Systematic error that occurs when PPTs do not remember previous events accurately or omit details.
What is selection bias?
Any type of bias that results in a sample population that is not representative of the population at large or population of interest or where results in study groups are systematically different from each other.
What are some types of selection bias?
3:
Allocation
Non-response bias
Sampling bias
What is sampling bias?
It occurs when the method used to sample the population means that some members of the intended population are more likely to be selected than others.
What is measurement bias?
Any systematic or non-random error that occurs in the collection of data in a study.
Also called information bias.
What are some types of measurement bias?
Recall bias
Observer bias
Instrument bias
Ascertainment bias
What is observer bias?
Systematic differences between what is observed and what is the truth.
What is confirmation bias?
When the overall research process is aimed at confirming the researcher’s perception or hypothesis about the research subjects.
What is social desirability bias?
This bias involves respondents answering questions in a way that they think will lead to being accepted and liked.